A. Galanin produced a profile similar to scopolamine and MK-801 on the delayed non-matching to sample memory task. Galanin decreased choice accuracy on DNMTS when microinjected into the lateral ventricle, ventral hippocampus, and nucleus basalis magnocellularis of rats. Galanin was found to stimulate feeding when microinjected into the paraventricular nucleus of the hypothalamus and central nucleus of the amygdala. New galanin antagonists, C7, M40, and M32, blocked galanin-induced feeding when administered intraventricularly or into the paraventricular nucleus of the hypothalamus, at doses of antagonists equimolar to or lower than galanin doses. B. Push-pull perfusion was found to have considerably higher recoveries than microdialysis, and allowed detection by standard radioimmunoassay of both cholecystokinin and galanin. Split samples from the nucleus accumbens allowed measurement of dopamine by HPLC and CCK by RIA in the same sampling period, in a feeding paradigm in rats. A CCK-B antagonist increased dopamine concentrations in microdialysate from the nucleus accumbens of anesthetized rats. 50 mM potassium stimulation increased galanin concentrations in push-pull perfusate from the ventral hippocampus of anesthetized rats.